DE102011085107B4 - Method for controlling a wind energy plant - Google Patents

Abstract

The present invention relates to a method for controlling a wind power plant, comprising the steps of detecting an internal generated in the first wind turbine, a fault of the first wind turbine fault signal indicating, receiving at least one external, generated outside the first wind turbine, a fault of another wind turbine fault signal indicating, evaluating the internal error signal in response to the at least one external error signal.

Description

The present invention relates to a method for controlling a wind turbine as well as a method for controlling a plurality of wind turbines. Furthermore, the invention relates to a wind turbine and a wind farm with multiple wind turbines.

Wind turbines are well known, they convert in today's conventional design kinetic energy of the wind in a kinetic energy of an aerodynamic rotor. The kinetic energy is converted by means of a generator into electrical energy or in electrical power when a snapshot is considered. A common wind turbine is in the 1 shown.

When controlling a wind turbine, care must also be taken that any disturbances in the wind turbine are met in order to exclude or reduce consequential damage to the wind turbine and / or a hazard, in particular a danger to persons. For this purpose, any disturbances in the wind turbine are monitored in order to initiate protective measures if necessary.

For example, a fire in the nacelle of the wind turbine can be detected by a smoke sensor. The wind turbine can then be stopped and, if necessary, deletions can be initiated. If an optical smoke detector is used, it can lead to a false alarm if the visibility is poor for reasons other than smoke, such as dust or ice crystals.

Another example is a noise monitoring of the wind turbine, which may be provided, for example, on the spinner, ie in the region of the cover of the rotor hub. By such a noise monitoring unusually loud noises can be detected, which may indicate an existing or imminent damage, in particular mechanical malfunction. In this case, the wind turbine - which is synonymous simply synonymous simply referred to as an attachment - stopped and contacted service personnel. Again, there may be a false alarm, for example, when a heavy hail shower on the wind turbine, including the spinner, down and thereby leads to an unusually high noise level.

These are just two examples of monitoring in the wind turbine, which can lead to a false alarm. If there is such a false alarm, a loss of income due to unwanted stopping of the wind turbine can be the result. On the other hand, a much greater damage to a wind turbine or by a wind turbine can occur if the mentioned or other warning signals are ignored while no false alarm.

The American patent application US 2009/0110539 A1 describes a wind farm with multiple wind turbines and an ice sensor for detecting an ice condition of at least one of the wind turbines. A main controller is connected there with the ice sensor and at least two wind turbines.

The German patent DE 10 2009 004 385 B4 relates to a method and an arrangement for monitoring a wind turbine. There, condition data of the wind turbine are recorded, deviations from a comparison value determined, a probability of failure and an availability value are determined.

The German patent application DE 10 2008 046 156 A1 relates to a method for diagnostic monitoring. It is described there that, on the system side, each status code is assigned operating data for evaluating the status code in order then to transmit a complex status signal to remote monitoring units.

The German patent DE 101 15 267 C2 relates to a method for monitoring a wind turbine. Accordingly, in the interior of the wind power plant, where electrical or mechanical units are used, the sound produced there is picked up by means of a sound pickup. Further, an optical pickup is provided, which is directed to the location resulting sound or is alignable.

The German patent DE 199 48 194 C2 relates to a method of monitoring wind turbines, wherein a reference noise spectrum of a wind turbine is recorded at at least a certain point of the system, this reference spectrum is stored and an operating noise spectrum is recorded during operation, wherein the recorded operating noise spectrum is compared with the stored reference spectrum, deviations capture.

The invention is therefore based on the object to address at least one of the above problems, in particular to reduce. In particular, a solution is to be created to detect error signals more reliably and to avoid false alarms without correctly generated error signals ignore or overlook. At least an alternative design should be created.

According to the invention, a method for controlling a wind turbine according to claim 1 is proposed. Accordingly, an error signal generated internally in the relevant wind turbine to be controlled is detected. This error signal indicates a malfunction of the wind turbine, such as the suspicion of a fire in the wind turbine or the suspicion of damage to a mechanical element. Hereinafter, in particular these two cases, namely an error message relating to a fire or error signals on the one hand and an indicated by a loud noise damage error signals or error messages are used for explanation. Nevertheless, the invention is not limited to these two basic applications.

The fact that the error signal, which can also be referred to as an error message in principle, indicates a malfunction of the wind turbine, is to be understood as meaning that a corresponding error criterion exists, such as, for example, the response of a smoke detector or the response of a noise sensor. The fact that the error signal indicates the relevant fault also includes the possibility that the fault is actually not present. The error signal may be the result of one or more sensors, or other states, such as operating conditions, may be evaluated. Likewise, multiple error signals may be present for different disturbances simultaneously.

Furthermore, an external error signal is received at the wind turbine. An external error signal is one that indicates a failure of another wind turbine. In particular, such an external error signal, which refers to the disturbance of an external, so another, wind turbine, in this other wind turbine an internal error signal, which is also transmitted to this first wind turbine. As a result of this transmission, it becomes an external error signal for the first wind energy plant.

The internal error signal is then evaluated by the first wind energy plant as a function of the at least one external error signal. The evaluation of the internal error signal is thus not limited to this internal error signal, but also takes into account at least one external error signal, namely namely one that affects the same disorder or disorder as the internal error signal to derive only from whether there is a fault or not.

Preferably, a protective measure for protecting the wind turbine, in particular the stopping of the wind turbine, depending on the internal error signal and also depending on the at least one external error signal is initiated. The initiation of a protective measure is thus not limited to the evaluation of the internal error signal. Thus, the one or more considered external error signals can then optionally decide whether it is assumed that a fault or not. In addition to stopping the wind turbine other protective measures come into consideration, such as the triggering of a deletion, if the error signal relates to a fire, to name just another example.

According to one embodiment, it is proposed that, in evaluating the internal error signal, it be assumed that there is no malfunction, when the internal error signal indicates a malfunction and at least one further external error signal also indicates a similar malfunction, namely a malfunction of the respective wind turbine, with the said external one Error signal is an internal error signal. If, for example, an internal error signal appears in the first wind energy plant that indicates a fire, in particular triggered by a smoke detector in the first wind energy plant, it is initially assumed that there is a fire. Is now an external error signal another, z. B. second, wind turbine, which also indicates a fire, namely a fire in the second wind turbine, this may be an indication of a false alarm. In fact, according to the invention, it has been recognized that it is extremely unlikely that a fire will occur in two wind turbines at the same time. Rather, in such a case, it can be assumed that the smoke detectors which have each issued the fire warning signal as an error signal have triggered a situation other than a fire, namely by circumstances which experience has shown and / or usefully to expect for several wind turbines at the same time , This is the case, for example, with a sandstorm. If there is a sandstorm, the visibility conditions can usually worsen at several wind turbines set up in the immediate vicinity of each other.

Such a phenomenon may also occur for other disturbances, such as noise monitoring. If an internal error signal indicates too high a noise level in the first wind energy plant, this may indicate a mechanical problem such as bearing damage. The noise sensors are in the first place or the usual way to detect loose parts and thus sporadically occurring Noises designed. But reports at almost the same time an external error signal, such as a second wind turbine, also too high a noise level, it is unlikely that both wind turbines have at the same time a mechanical, a loud noise generating damage. In particular, it is unlikely that a part will be released in both wind turbines at the same time. Instead, the simultaneous occurrence of a loud noise at several wind turbines arranged close to each other may rather indicate, for example, a hailstorm. A hail event usually does not occur in isolation on a single wind turbine but, for example, almost simultaneously in an entire wind farm.

Thus, it comes to the first surprising finding that the presence of several similar error signals or error signals for several similar disorders, but rather indicates that such a disorder does not exist then. This knowledge is taken into account by the described method and implemented accordingly.

According to one embodiment, a method is proposed, which is characterized in that when the internal error signal indicates a fault of the first wind turbine, a protective measure, in particular the stopping of the wind turbine is initiated when the received external error signal or first one or several of the received external error signals, not indicative of a similar disturbance of the other wind turbine concerned. If, however, the received external error signal or one of the received external error signals indicates a similar disturbance of the respective other wind turbine, a warning, in particular a warning signal, is generated without a protective measure being initiated. As explained, the additional evaluation of at least one external error signal may indicate that, despite the presence of an internal error signal, the disturbance thus indicated is unlikely. However, the said disturbance is nevertheless not excluded, and by providing a warning signal in this case, it is possible to further analyze the respective underlying situation in detail.

An analysis of such a situation - which may also be carried out without the presence of a warning signal - can for example be automated by the wind turbine or its control or a central control. In this case, for example, further sensor data, such as the temperature of a temperature sensor, evaluated. Additionally or alternatively, an analysis can be done manually. In particular, due to known weather or environmental conditions or from other empirical values out a reason for the internal error signal can be found to ensure that the displayed fault is not present. Likewise, the analysis may reveal that the displayed disorder, contrary to expectations, is present.

Preferably, such a warning signal is sent to or generated by a control center. In this case, the warning signal may preferably include information about the error signals, namely the internal or a plurality of internal error signals or the one or more external error signals. This information can be used accordingly to find out the actual cause of the respective error signals.

Preferably, a so-called SCADA is used, the name of which originates from English and abbreviates "Supervisory Control And Data Acquisition", and is nevertheless familiar in the German language. In wind turbines, such a data system is used to transmit measurement and control data between individual wind turbines and also to a control system. The corresponding error signals can be provided on this SCADA system. Thus, any wind turbine that participates appropriately in this system provides its internal error signals as external error signals on the SCADA. Preferably, in this case, the respective error signal includes not only its nature but also its origin, namely in particular from which wind turbine it originates. As a result, the error signals, in particular also adjacent wind turbines, in particular a wind farm, are available and each wind turbine involved can in its own way include the information provided, in particular the error signals provided in a separate error signal evaluation basically at will.

According to a further embodiment, it is proposed that the internal error signal and / or a relevant sensor signal be suppressed for a predetermined time if the internal error signal indicates a fault and at least one external error signal respectively a similar fault of the respective other wind turbine or one of the respective indicates other wind turbines.

It is advantageous to provide a comprehensive method that is used to control multiple wind turbines. At least one of the wind turbines is operated by a method of at least one of the mentioned embodiments. Preferably, all of these total control underlying wind turbines each controlled by a method according to one of the described embodiments. Preferably, the wind turbines exchange their error signals with each other in order to make the estimates described above. This can be done directly or via a control center such as a SCADA. In particular, at least one of the wind turbines provides at least one internal error signal to the other wind turbines as an external error signal. Accordingly, there is an advantageous overall concept for controlling these wind turbines.

According to the invention, a wind turbine is also proposed, which has an aerodynamic rotor for generating a rotational movement from wind and also has an electric generator for generating electrical power from the rotational movement. The wind turbine employs a method of one of the illustrated embodiments for control. Such a method can be implemented in the wind energy plant, for example, on a control device and / or a process computer. Accordingly, the wind turbine is prepared to evaluate a fault more difficult by not only an internal error signal is taken into account, but other external error signals or at least one of them, which are generated and provided by other wind turbines.

Preferably, a wind farm is proposed, which has a plurality of wind turbines, at least one wind turbine is designed as described above and is operated in particular by a method of the described embodiments. The wind farm is prepared to use an overall, overarching process whereby all or at least part of the wind turbines of the park benefit from the additional information of the other wind turbines in the park. Preferably, the wind turbines are coupled to each other via a SCADA system for exchanging data. It is advantageous if a central control center is provided which manages and / or forwards central data. A wind farm is characterized in particular by the fact that it uses a feeding node common to all wind turbines for feeding electric power into an electric grid.

With the solutions described, it may be possible to prevent the initiation of a protective measure when an internal error signal occurs if at least one other error signal from another wind farm indicates a similar fault. It may be that the one wind turbine, in which this error signal occurs first, emanating from the disturbance indicated thereby and initiates a corresponding protective measure. Maybe this wind turbine will be stopped unnecessarily. Only when - possibly a short time later - further error signals from other wind turbines occur, the assumption is suggested that the displayed fault does not exist. It can now be decided depending on the situation, whether the already stopped wind turbine - to stay in this example - starts again or for safety's sake stops until the error signal is clarified.

Basically, it is proposed to question the statement of an error signal about an indicated disturbance based on statistical considerations. Preferably, it is proposed to consult not only a single external error signal, but to consider at least a second error signal and to preclude the presence of the fault indicated by the internal error signal only if at least two external error signals indicate the same fault. Preferably, the presence of a displayed fault is precluded only when three external fault signals or more preferably at least four external fault signals indicate a similar fault in the respective other wind turbines.

According to a further or alternative embodiment, error signals are taken into a causal relationship between a plurality of wind turbines. In particular, this includes error signals relating to situations resulting from the local proximity of several wind turbines. This includes in particular an ice accumulation detection. When several wind turbines in a wind farm usually very similar weather conditions are based. If several wind turbines in a wind farm detect ice accumulation, there is a high probability that there will be an ice accumulation in the wind farm altogether. Such a conclusion can also depend on how much and which wind turbines in the wind farm have recognized an ice accumulation and ultimately it can also depend on how precise the Eisansatzerkennung the respective wind turbine. For example, if ten out of eleven wind turbines in a wind farm have identified ice accumulation, the likelihood is high that the eleventh wind turbine will also have ice accumulation. However, this is not true if this eleventh wind turbine either has such a reliable Eisansatzerkennung that you can rely on 100% ig, or speak for other reasons concretely in this system against an ice accumulation, such as a more secure location, or already in operation taken heating of the rotor blades to prevent an ice accumulation.

It is preferably proposed to evaluate the associated error signals of several wind turbines, in particular to evaluate a plurality of wind turbines of a wind farm, for such errors that are causally related via a plurality of wind turbines. Such an evaluation preferably includes the evaluation of how many of the investigated wind turbines, in particular how many of the wind turbines of the wind farm currently output this error signal. In addition, it is proposed to evaluate further details of the wind turbines that have issued such an error signal, in particular their location in the wind farm. Such an evaluation can be made wholly or partly on a wind turbine, on all wind turbines or performed on a central process computer. The result is preferably output as a parking disturbance signal and may, for example, indicate the number of wind turbines that have such an error signal, in particular in relation to all wind turbines of the wind farm. This information can be given as an absolute value or as a percentage of all existing wind turbines of the wind farm. Preferably, the parking disturbance signal is output as a probability or as a frequency for said error case in the wind farm, or it is used to calculate such probability or frequency.

In that regard, it is proposed to differentiate for the same errors of different wind turbines between those causally related between these wind turbines, and those that are not causally related. If the same error signals of different wind turbines occur for faults which are not causally connected, it is assumed that corresponding fault signals were erroneously output or that they were likely to have been output by mistake. Accordingly, it may be possible to justify ignoring the error signal.

For several fault signals of different wind turbines, which indicate a same, causally related error, it is assumed that at least one probability can be derived therefrom that the fault is present in still further wind turbines, possibly in all wind turbines of a wind farm.

Such a causal error relates in particular to an ice formation and correspondingly to an ice accumulation recognition or ice recognition. Preferably, it is thus proposed that if one or more plants have already been stopped in a wind farm with a corresponding ice detection status, it can be achieved by means of a corresponding function, which can be referred to as park ice detection, that all other plants of the park or selected plants of the park Parks are stopped, even if they themselves have not recognized any ice accumulation.

The wind turbines exchange information with each other via SCADA. Each status of an ice accumulation is transmitted by SCADA once at the time the status is created, together with the number of systems in the park and the number of the system that generated the status, to all systems in the park. Each facility stores this information in a spreadsheet of all other facilities in the park and calculates a so-called percentage of park icing with each change in the park's facilities and the number of sites that have detected ice accumulation.

If this degree of park iciness, which can also form the parking disturbance signal, is greater than a predetermined value, in particular greater than a value that can be specified in the control software, the wind energy plant, which is also referred to as plant for the sake of simplicity, is stopped with an ice accumulation detection status. However, this status, since it was not generated at the relevant plant, is not distributed to all plants and thus does not affect the calculated degree of park icing.

Preferably, the systems that have generated an ice detection status send a notification to all other facilities as soon as the ice bank detection has been reset. Such a status message is transmitted to all plants in the wind farm using SCADA. Each system then deletes the ice accumulation detection of the plant, which has generated the information that there is no longer any trace, from its table and recalculates the degree of park icing. If this degree of park ic is then lower than the set limit value, the system restarts.

For example, if a limit of 20% is set, this means z. For example, in a park with ten plants, this plant will be stopped when two or more plants generate an ice detection status. To disable the park detection, a value of 100% must be set for this limit.

Preferably, such values can also be manually deleted or reset to 0.

A reset of such a Parkeiserkennung should preferably be carried out only in exceptional cases. To plants, which were stopped by such a park recognition again in Instead of resetting the park detection, ice detection should generally be reset on sites that were stopped due to ice build-up. As a result, the degree of park icing can fall below the limit, so that the wind energy plants stopped depending on this limit value can restart.

Likewise, a single asset that has been stopped due to park detection can be re-started - even if only for a short time for maintenance - by setting the limit to 100%, effectively disabling park detection for that asset.

The invention will be explained in more detail by way of example with reference to specific embodiments with reference to the accompanying figures.

1 shows a wind turbine using a method according to the invention.

2 shows schematically three interconnected via a SCADA system wind turbines for the exchange and joint evaluation of error signals.

3 shows a flowchart of an embodiment for controlling a wind turbine.

1 shows a wind turbine 100 with a tower 102 and a gondola 104 , At the gondola 104 is a rotor 106 with three rotor blades 108 and a spinner 110 arranged. The rotor 106 is set in operation by the wind in a rotary motion and thereby drives a generator in the nacelle 104 at.

2 schematically shows three wind turbines 2 that have a data management system 4 are connected, which is conducted here wired, but can also be provided as a radio link. This data management system also connects to a so-called SCADA system 6 ago. The SCADA system 6 includes a statistics block B illustrated herein. Information such as an error signal for indicating a fault is provided by the wind turbines, respectively 2 to the SCADA system 6 directed. This is indicated by the information arrow 10 each schematically indicated. The collective arrows 12 illustrate that the information in the SCADA system 6 converge. From the information can by means of the statistics block 8th a statistical evaluation will be made. This includes evaluating error signals from multiple wind turbines together to determine if these error signals are actually indicative of a fault or if it may be a false alarm. Here, for example, from the wind energy plant shown on the left 2 an error signal as internal to this wind turbine error signal together with error signals of the other two wind turbines 2 evaluated as external error signals. The presentation of the 2 is held so that the three illustrated wind turbines 2 functionally equally represented. Thus, depending on the situation, namely, depending on where an error signal first occurs, each of the wind turbines 2 be the first wind turbine in the sense of the above explanations, which provides an internal error signal. In the example shown, the remaining two wind turbines may each be the other wind turbines, each supplying an external error signal.

The evaluation is in 2 as shown in the SCADA system 6 is carried out. Likewise, the evaluation can be carried out in each case in the wind turbine. In any case, a SCADA system is used at least for information forwarding, preferably also for information evaluation.

2 also illustrates a storm cloud 14 , When this storm cloud 14 leading to a thunderbolt with accompanying thunder, this thunder can produce such a loud noise in all the illustrated wind turbines 2 a noise sensor triggers in any case, so in each of the wind turbines 2 , generates a corresponding error signal. The dashed lines of action 16 should illustrate that the thunder on each of the wind turbines 2 effect. In this example shown, each of the wind turbines then generates 2 a corresponding error signal, so that a total of three error signals generated and in the SCADA system 6 be evaluated together. In this illustrated example, if there is an impact, then at the same time there are three wind turbines 2 each have generated a caused by a loud noise error signal. From these it is concluded in these examples that this can not be a plant defect like a loose part causing a rumble, but one for all three wind turbines 2 concurrently effective phenomenon. By additionally evaluating a weather report, be it automated or manually by a person, the suspicion can be substantiated that there is no disturbance and rather a gathering thunderstorm must be the reason for the triggering of the noise sensors.

It is based on the realization that it is unlikely that all three wind turbines 2 at the same time have a technical damage that causes the noise sensors. Even if one of these three wind turbines 2 the noise sensor strikes first, because the thunder, for example, closer to this system is tuned possibly these shut down for safety's sake, but after receiving further error signals results from probability considerations that not everywhere such an error may be present, and accordingly not all wind turbines are switched off and must not be restarted manually accordingly. This can increase their availability. Following this situation, it is also possible, if appropriate, to restart the initially switched off energy system. This can be done manually or automatically depending on the embodiment.

This thunder with any noise-motivating error signals is just one example. Other phenomena and other error signals can be treated accordingly.

3 FIG. 1 schematically illustrates a flowchart of a proposed score according to one embodiment. Accordingly, in the block 30 recorded an internal error signal (F _int ) of a first wind turbine. The blocks 32 and 34 illustrate that basically any number of further external error signals (F _{ex, 1} to F _{ex, n} ) can be recorded. Basically, this depends on the number of available wind turbines. The example of 3 thus concerns n + 1 wind turbines, namely the first and n more.

The error signals, both the internal error signal and the external ones, are in the evaluation block 36 merged. It should be noted that in the example shown, an error signal is assumed which is always present but whose value indicates first whether there is a fault or not. For example, the error signals can each assume the value zero for no disturbance and 1 for accepting a fault. Alternatively, a transmission of an error signal takes place only when an error occurs.

The merged information is then first in the query block 38 evaluated for the internal error signal F _int . If the evaluation is negative, so there is no error signal for the first wind turbine, so everything is OK and the flow branches to the first output block 40 who breaks off the evaluation in any case for the first wind turbine at this point and optionally outputs a o.K. signal. However, the output of the signal is not important, as it is only in case of failure that a need for action arises.

Is the result of the first query block 38 positive, so there is an error signal for the first wind turbine, so there is another query in the second query block 42 , Here it is checked whether at least one of the external error signals F _{ex, 1} to F _{ex, n} also indicate a similar fault or a similar fault.

If this is not the case, then it is concluded that there is actually a fault in the first wind turbine, and the system is thus stopped, which by the protective block 44 is clarified. The system can be stopped accordingly, there are also other or additional measures into consideration.

Is in the second query block 42 the result is positive, so if there is at least one external fault signal indicating the same fault in another wind turbine as the internal fault signal of the first wind turbine, no protective measure is taken, at least not immediately, but instead a warning according to the warning block 46 output. This warning can be further processed in the SCADA, and it can be decided later, if it is based on further information is finally assumed by a false warning or if the first wind turbine must be stopped or other protective measures must be taken.

In the query block 42 can also be made a query so that at least two or at least three or at least four external error signals, ie according to many other wind turbines must display the same fault as the internal error signal to finally the warning block 46 and not to the protection block 44 to branch.

This in 3 In principle, the scheme shown can be carried out successively or in parallel for each wind energy installation of the considered wind energy plant collection, in particular of a corresponding wind farm.

Thus, according to the invention, the significance of any error signals is improved in a simple manner, in particular without additional hardware expenditure, via a statistical evaluation. Any unwanted plant shutdowns can be avoided.

Thus, statistically determined probabilities over several plants of a wind farm can be used to correct any incorrect measurements or incorrect evaluations. Downtimes of systems triggered by a false alarm from such a faulty measurement can be avoided.

The central SCADA system records all states of all wind farm plants. This required data can also be called status data. Selected states will be recorded statistically, namely in particular as number of installations with the same state at the same time, and sent back to the facilities as statistics. The self-sufficient system controls can thus avoid false alarms by including statistics. The SCADA system thus creates a data compilation and database setting, but the wind turbine itself is left to decide what consequences it must draw from this. According to one embodiment, stand-alone plant controls which are present in the plants can thus avoid false alarms or use the states of other plants and react accordingly to them.

It creates a cost-effective solution, since only a one-time software effort is required.

The following concrete examples are again pointed out for explanation:
In severe thunderstorms and hailstorms with extremely large hailstones, the state of the art leads to false triggering of the noise sensors. In order to remedy this situation, an information exchange of the facilities in the park is proposed with the help of SCADA.

As soon as an installation transmits to SCADA appropriate information, such as "noise detected in a spinner", which usually precedes a corresponding status ("noise in a spinner") 50:14, SCADA sends this information directly to all facilities in the park. Thus, each system has the ability to detect whether noise sensors from other systems in the park have responded within a certain period of time.

For example, if two systems in a park with two or three systems or three or more systems in larger parks detect noises within 30 minutes, the noise sensors of all systems in the park will be deactivated for at least 30 minutes. All systems generate status information "Noise sensor deactivated". 30 minutes after the last sound from a system has been detected or after a maximum of 5 hours, the sound sensors are reactivated. Here, instead of the aforementioned 30 minutes, as well as other than the aforementioned 5 hours, other time values can be used in principle. All systems then transmit the information "Sound sensor activated" to SCADA. This information can be used to determine when and for how long noise sensors were disabled.

If a system has already shut down due to a thunderstorm or a hail shower with the above status 50:14 before a sufficient number of other systems have also detected noises in the spinner, then the so-called reset lock is canceled in this system and the fault is acknowledged. The system will now automatically start up again with the noise sensor deactivated. The prerequisite for this, however, is that the status 50:14 is not yet available for more than 30 minutes. Then it must be assumed that there is another cause for the status.

The reset lock prevents the system from being put back into operation by the operator. The reset lock can be canceled after a service code has been entered by a service employee.

Due to thunderstorms caused false triggering should be excluded by these measures in the future as far as possible. For single systems, this proposed change in the control system is basically ineffective, so that a software change can generally be proposed in general, without having to be taken into account whether the system is to be installed as a single system or in a park. Accordingly, it may at most lead to occasional false tripping of individual systems during thunderstorms possibly having to be accepted.

As a further example, reference is made to previously existing problems with smoke detectors explanatory. The finest ice crystals and desert dust in the state of the art have repeatedly caused false triggering of smoke detectors in the nacelle at some locations. For this purpose, too, it is proposed, basically still with the noise sensors described above, that an information exchange of the installations in a park take place with the aid of SCADA.

As soon as an installation transmits the status "smoke detector (hatch)" or "smoke detector (spinner)" to SCADA, this information will be sent back or distributed by SCADA directly to all systems in the park. Thus, each system has the ability to detect whether smoke detectors of other systems in the park have responded within a certain period of time.

If smoke detectors respond within 5 hours to two systems in a park with two or three systems or to three or more systems in larger parks, the two smoke detectors on the hatch and spinner in all systems in the park will be deactivated for 24 hours. All systems then generate the information "Smoke detector: hatch and spinner disabled". At the end of the 24 hours, the smoke detectors are activated again with the information "Smoke detector: hatch and spinner activated" for at least 6 hours. These times can vary as well. This prevents several, for example, defective smoke detectors in a park from permanently deactivating the smoke detectors of all systems.

Since the deactivation of the smoke detectors preferably only takes place when two or three systems with the status "automatic deactivation of the smoke detectors" have already switched off, the reset lock is canceled again in this system and the fault is acknowledged. The system will now automatically start up again with the smoke detectors deactivated. However, the prerequisite for this is that the status "automatic deactivation of the smoke detector" is not yet available for more than 5 hours. Then it must be assumed that there is another cause for the status.

Claims (11)

Method for controlling a first wind energy plant ( 2 . 100 ), comprising - detecting an internal generated in the first wind turbine, a fault of the first wind turbine indicating fault signal (F _INT ), - receiving at least one external, generated outside the first wind turbine a fault of another wind turbine fault signal indicating (F _EX ), - evaluate the internal error signal (F _INT ) as a function of the at least one external error signal (F _EX ), wherein the evaluation of the internal error signal (F _INT ) is based on the absence of a fault, if the internal error signal (F _INT ) indicates a fault and the at least one external or at least one of the external error signals (F _EX ) each indicates a similar disturbance of the relevant external wind turbine. A method according to claim 1, characterized in that at least one protective measure for estimating the first wind turbine ( 2 . 100 ), in particular stopping the wind energy plant ( 2 . 100 ), depending on the internal error signal (F _INT ) and depending on the at least one external error signal (F _Ex ) is initiated. Method according to one of the preceding claims, characterized in that, if the internal error signal (F _INT ) is a fault of the first wind energy plant ( 2 . 100 ), - a protective measure, in particular stopping of the wind energy plant ( 2 . 100 ), is initiated, if the received external error signal (F _Ex ) or one of the received external error signals (F _EX ) does not indicate a similar disturbance of the relevant other wind energy, - a warning, in particular a warning signal is generated, if the received external Error signal (F _EX ) or one of the received external error signals (F _Ex ) indicates a similar disturbance of the other wind turbine concerned, in particular, then no protective measure is initiated. A method according to claim 3, characterized in that the or the warning signal which is generated when the internal error signal (F _INT ) indicates a fault and the at least one external or at least one of the external error signals (F _EX ) each have a similar disorder the relevant external wind turbine, is sent to a control center and information of the internal and at least one external error signal (F _EX ) includes. Method according to Claim 3 or 4, characterized in that the or a warning signal and / or a transmission or provision of the at least one external error signal (F _EX ) takes place by means of SCADA. Method according to one of the preceding claims, characterized in that the internal error signal (F _INT ) and / or a respective sensor signal for a predetermined time is suppressed when the internal error signal (F _INT ) indicates a fault and the at least one external or at least one of the external error signals (F _Ex ) in each case indicates a similar disturbance of the relevant wind turbine. A method according to claim 1 or 2, characterized in that when evaluating the internal error signal (F _INT ) is assumed by the presence of a fault when the internal error signal (F _INT ) indicates a fault and the at least one external or one of the external error signals (F _Ex ) each indicates a similar disturbance of the respective external wind turbine, in particular an ice accumulation, in which case a parking disturbance signal is generated which indicates to further wind turbines that there is a disturbance concerning other wind turbines or if it is likely to affect further wind turbines , which have not generated a corresponding error signal itself. Method for controlling several Wind turbines ( 2 . 100 ), wherein at least one of the wind turbines ( 2 . 100 ) is controlled by a method according to one of the preceding claims and at least one internal error signal (F _INT ) of at least one of the other wind energy plants ( 2 . 100 ) is provided as an external error signal (F _Ex ), in particular that the wind energy plants ( 2 . 100 ) exchange their error signals with each other. A method according to claim 8, characterized in that in the presence of an error signal to an error condition, in particular Eisansatzerkennung, at least two wind turbines another, in particular adjacent wind turbine, which has not generated such an error signal, is placed in the same error state as the wind turbines, which have detected the fault, and / or generates a parking disturbance signal and is transmitted to at least one or the wind turbine, which has not generated an error signal of said error. Wind energy analysis ( 2 . 100 ) with an aerodynamic rotor ( 106 ) for generating a rotational movement from wind, with an electric generator for generating electrical power from the rotational movement, wherein the wind energy plant ( 2 . 100 ) is prepared for controlling by means of a method according to one of claims 1 to 7, in particular that the wind energy plant ( 2 . 100 ) has a control device on which a method according to any one of claims 1 to 7 is implemented. Wind farm with several wind turbines ( 2 . 100 ) and at least one wind turbine ( 2 . 100 ) according to claim 10, wherein the wind farm is prepared to carry out a method according to claim 8 or 9.

Images